criminal law,

Privilege against Self-Incrimination in Civil Injunctions: the Double-Edged Sword of Themis?

By Jason Lin Oct 09, 2019

In general, the privilege against self-incrimination (“PSI”) protects a defendant from being compelled to provide (potentially) incriminating statement or information. PSI may be asserted by a defendant in a civil proceeding in opposition to a disclosure order. Like the double-edged sword of Themis, PSI cuts both ways. Upholding a defendant’s assertion of PSI may damage the interest of a claimant. Abrogating PSI may benefit the claimant at the cost of a defendant’s right to trial.  This article argues that the respective approaches of courts in Australia and England & Wales are both unsatisfactory. The main aim of this article is to persuade readers that this area of common law must be replaced by a statutory abrogation of PSI and corresponding prohibition against the use of information obtained in such disclosure order against the defendant in a subsequent criminal prosecution.  


The exact origin of this common law right in the United Kingdom is outside of my repertoire. Briefly, it may date back to the abolition of the Court of Star Chamber, which was notorious for extracting admissions of guilt using torture.1 It more likely evolved from the common law’s repackaging of the Canon law doctrine of nemo tenetur seipsum accusare (no one is bound to incriminate themselves).2

The United States of America enshrined this common law right under the 5^th^ Amendment of its constitution. Justice Goldberg in the Murphy v Waterfront Commission3 case provides a powerful rationale for the PSI rule based on the principles of liberty and fair trial:

“The privilege against self-incrimination ‘registers an important advance in the development of our liberty – one of the great landmarks in man’s struggle to make himself civilized…It reflects many of our fundamental values and most noble aspirations: our unwillingness to subject those suspected of crime to the cruel trilemma of self-accusation, perjury or contempt…“4

A defendant in a civil trial is also entitled to the protection of PSI. Indeed, it would be quite inconsistent if the law protects defendants in criminal proceedings but not in civil proceedings, for what is said in a civil court (under oath/ affirmation) may be equally prejudicial and incriminating. For instance, it is not uncommon for courts to grant an ancillary disclosure order in aid of a Mareva5 injunction.6 The rationale is simple: the effectiveness of the asset-freezing order would be in serious doubt if the claimant has no idea where the defendant’s assets are located, such that it is very hard to monitor whether the order is complied with. The defendant is bound to disclose the information identified in the order against the threat of contempt. Here, if PSI does not apply in civil trial, the defendant is forced to either incriminate himself or perjure. The effect would be that PSI is infringed categorically in civil proceedings: an unthinkable result.

Criticism of PSI in Civil Trials

Despite the real threat to a defendant’s right to fair trial, PSI in civil trials is subjected to formidable judicial criticism. As noted by Lord Neuberger MR in Gray, Lord Templeman once described PSI as “an archaic and unjustifiable survival from the past” and questioned whether PSI should be retained in civil proceedings at all.7 The reason given for such a strongly worded criticism is that PSI in civil trials does nothing to protect a defendant from being ill-treated by police or prevent confession wrongly extracted from being admitted in a criminal court, as structural reforms of the way police treats and interviews suspects are needed to ensure fair treatment of criminal defendants.8 For instance, the prosecution may adduce evidence of confessions improperly obtained. If admitted, such confession would not be subjected to protection by PSI. However, Lord Templeman’s objection cannot be put too highly. Protecting a defendant from ill treatment of police is one of the many concerns of a criminal justice system. A defendant may be compelled by the courts on application by the claimant to disclose incriminating/ potentially incriminating information about themselves. This has nothing to do with the police treatment of suspects.

The author nonetheless agrees with Lord Templeman to the extent that PSI in civil trials should be treated with suspicion. Besides the reasons given by Lord Templeman, it is submitted that there are at least two other reasons for not categorically upholding PSI in civil proceedings. Firstly, there is a fundamental difference between the subject matter of a civil and criminal trial. The prosecution, if successful, will result in criminal penalties and related social/ political costs to the defendant. For instance, the defendant may be imprisoned and/ or have their future job prospect and social standing affected. In contrast, the claimant is attempting to obtain civil remedies, largely in the form of monetary awards. Thus, a civil defendant who is forced by court to make disclosures is facing less serious consequences in comparison to a criminal defendant who is forced to answer incriminating questions.

Of course, the first reason does not lose sight of the possibility that what’s said or disclosed in a civil case could be used in a criminal case (as noted above). Thus, the second reason looks at the role of alternative protection in criminal and civil trials. In a criminal trial, it makes no sense to talk about alternative protection given in substitution to PSI. The purpose of a criminal trial is to determine whether the defendant committed the crime(s) charged. What is said or adduced in evidence goes to that purpose. It would make a laughing stock out of courts to force the defendant to make incriminating statements while telling the jury (or the judge reminding her/himself) to disregard such statement. However, in a civil trial, alternative protection could potentially work to achieve a balance between protecting the defendant and upholding the interest of the claimant (discussed in more detail below).  

However, the author disagrees with Lord Templeman in that PSI should not be banished altogether from civil courts. Indeed, even Lord Neuberger MR qualified his criticism of the PSI by saying that it can be abolished if adequate alternative protection along the lines of section 72(3) of the Senior Courts Act 1981 (discussed below) could be offered.9 Lord Neuberger MR’s cautious approach is laudable for three reasons. Firstly, the risk of prejudice to the claimant is comparatively less serious than that to the defendant. A claimant may receive a less effective form of civil remedy on the one hand. On the other hand, the risk to a defendant’s liberty and right to fair trial is arguably graver. Secondly, in interlocutory proceedings, once disclosure is made, information cannot be ‘un-disclosed’ if it turns out to be wrongly granted after substantive hearing. The damage to the defendant’s right to fair trial is permanent. Thirdly, the grant of interlocutory injunction is not subjected to a high bar (see the American Cyanamid10 principle). This means that defendants may be compelled to incriminate themselves without the claimant fully proving its case.

If the right to assert PSI in civil proceedings should be subjected to limitations to strike a balance between protecting the defendant and ensuring a claimant gets the just remedies, two further issues arise: (1) What is sufficient alternative protection? And (2) Is statutory reform the way forward?

Abrogating PSI and the adequacy of protection question

The English & Welsh line of authorities suggest that courts may abrogate PSI if it is satisfied that adequate protection could be substituted. However, this has not always been the case. In Bishopgate Investmnet Management Ltd, Dillon LJ stated very strongly that “any decision” to limit or curtail PSI is “essentially a political decision…[which is] a matter for Parliament”11. However, subsequent cases instead ask whether sufficient alternative protection could be given to a defendant. See for instance in AT&T Istel Ltd, the court was satisfied the defendant could be adequately protected, hence a disclosure order was made. In that case, the CPS submitted a letter acknowledge that it would not be using any information disclosed pursuant to a disclosure order. The judge thus saw fit to issue an order forbidding the use of disclosed information in any subsequent criminal prosecution. While this raises an interesting constitutional question as to the limits of judicial power to modify long standing common law right, it is better dealt with in more depth and focus in a separate piece.

The current English & Welsh approach begs the question however, what is sufficient protection. One may think it could take the form of a claimant’s undertaking to the court that it or any of its associates shall not disclose the information to the police or prosecuting authority. It may seem that such an undertaking is sufficient. Afterall, if a claimant discloses such information, (s)he may be liable for contempt.

However, such undertaking misses the point of PSI. The aim of PSI is to protect a defendant against the risk of prosecution and to prevent compelled statements to be relied on in a criminal trial. The undertaking merely prevents a claimant from adducing the disclosed information to the law enforcement authorities, it does not prevent any such evidence to be relied on in a court by a prosecutor. While the question arise as to the possibility that the prosecution could obtain such evidence, the existence of a possibility cannot be ignored and that the mere fact of disclosure may “set in train a process which may lead to incrimination or may lead to the discovery of real evidence of an incriminating character.”12

Indeed, a line of English authority suggest quite correctly that on top of an undertaking by the claimant, there must also be an undertaking from the prosecution. In Rank Film, Lord Wilberforce stated in the context of an Anton Piller13 order that the protection offered by an undertaking from the plaintiff is only partial, since it does not bind the prosecution.14 Thus,  the Court of Appeal in United Norwest Co-Operatives Ltd15 allowed an appeal against a disclosure order on the ground that the defendant’s right to assert PSI was infringed, as the Crown Prosecution Service (“CPS”) was not notified of the proceeding and no undertaking was offered. The reason why a CPS undertaking is required is that the CPS cannot be bound by a court order against its will in making its prosecutorial decisions. However, the CPS may, as it did in AT&T Istel LTd, undertake not to use material disclosed pursuant to a disclosure order against the defendant in a criminal proceeding.16 However, without such undertaking, the court would not be able to provide an adequate substituted protection.17

The Australian approach affords comparatively stronger protection to the defendant. In Reid v Howard18, the Court of Appeal of New South Wales followed the House of Lords’ decision in AT&T Istel Ltd19 by Handley JA. However, the Court of Appeal decision was reversed by the High Court of Australia. In the majority judgment, the PSI “is not simply a rule of evidence, but a basic and substantive common law right” and can only be overridden by legislation.20 The High Court’s decision in Reid v Howard case is followed in subsequent cases. In Vasil, Fitzgerald JA unequivocally stated that an order of disclosure will be set aside if it infringes PSI without statutory warrant.21 Clearly then, the Australian authorities are in direct conflict with the English position. It will be argued below that both the Australian approach and English approach are problematic.


It is submitted that the Australian courts’ refusal to abrogate or limit PSI even where adequate substitute protection is available is problematic. There is a paradox here: if the source of PSI is common law, the courts should be able to qualify or even abolish it. Unfortunately, the justices in Reid did not discuss this problem. The rationale seems to be that PSI is “deeply ingrained in the common law” and “lies at the heart of the administration of the criminal law”.22 It is not very clear whether the justices were hinting that the PSI has developed beyond a common law principle, such that it is part and parcel of the whole justice system above the reach of courts.

Nevertheless, the Australian position here is questionable at the minimum. It takes an extremely cautious approach: whenever a defendant shows PSI is engaged, a disclosure order would be refused. Is this level of caution necessary to protect the interests of a defendant? If it can be shown that the prosecution will be precluded from relying on information disclosed pursuant to a civil court order, then there should not be a risk to a defendant’s right to fair (criminal) trial. In fact, PSI is not ignored. The only reason why a court would go a step further in precluding the use of disclosed information in a subsequent criminal case is because it had the defendant’s PSI in mind. In such a case, the rationale for upholding PSI seems to drift away, so does the legal reasons for depriving the claimant of its rightful remedies. It can therefore be said that the High Court of Australia failed to pragmatically consider the practical impact of a disclosure order on the defendant’s PSI.


The E&W approach involves an undertaking from the CPS, which immediately attracts objections based on practicality. The central criticism here is twofold: the first is legal certainty. It is hard to see what incentive the CPS has in giving such undertaking. In fact, there are plenty of disincentives for giving such undertaking, as it may irreparably damage an ongoing investigation or prosecution, on top of increasing their administrative workload. Further, as noted by Professor Neil Andrews, justice is not always done in every case, since it remains the prosecutor’s discretion to order disclosure even if all the legal requirements are satisfied. It is hardly acceptable that whether or not a claimant gets relief depends on the discretion of the prosecutors.23 The author agrees with Professor Andrews, consistency is the benefit to both parties here. The defendant should have its basic right protected consistently and not live the fear that the CPS may exercise the discretion against them. Relief for the claimant should also be certain where the legal test for it is satisfied.

The second, more jurisprudential objection, is that the CPS should not be the ultimate arbiter of whether a claimant receives legal redress. The CPS’ decision to give an undertaking will likely be based on how ‘badly’ they want the defendant. This is a consideration alien to the civil dispute between the defendant and claimant. It is outside of the control, and often unknown to both parties. One may say this is like the judge’s consideration of public or legal policy. It is not. The CPS puts its prosecutorial interest first and not the interest of civil justice. The claimant and defendant can have a reasonable idea of the sort of legal/ public policy the judiciary is keen to uphold. The parties cannot know the minds of prosecutors.

In effect then, it may be argued that while there is a jurisprudential basis between the approach adopted by Australian and UK courts, there is no big practical difference in terms of their application. Recall that in AT&T Istel Ltd, the majority reiterated an obvious point: that the CPS cannot be bound by a court order Against its will.24 As the CPS is unlikely to give such undertaking (noted there aren’t any comprehensive survey or study on this), the current approach of the E&W courts means most claimants will win an injunction but practically an ineffective one.

Justice and fairness

The current English approach may also be criticized for standing justice on its own head when a civil court makes an order that’s binding on the criminal court. The majority in Reid v Howard stated that “it is inimical to the administration of justice for a civil court to compel self-incriminatory disclosure, while fashioning orders to prevent the use of the information thus obtained in a court vested with criminal jurisdiction with respect to the matters disclosed”.25 The majority objects to an order preventing the prosecutor from using such disclosed evidence in criminal prosecution. Lord Wilberforce stated a different concern in Rank Film: “I cannot accept that a civil court has any power to decide in a manner which would bind a criminal court that evidence of any kind is admissible or inadmissible in that court.”26 His Lordship’s objection is directed at a civil court binding a criminal court. To assess the strength of both objections, it is proposed that one look at AT&T Istel Ltd.

In AT&T Istel Ltd, the High Court fashioned an order preventing the use of the information disclosed from being used in criminal prosecution of the defendant. The exact words used are:

“No disclosure made in compliance with paragraphs 18 to 32 inclusive of this order shall be used as evidence in the prosecution of the offence alleged to have been committed by the person required to make that disclosure or by any spouse of that person.”27

 Arguably, the wording does not compel the criminal court into rejecting such evidence, but rather it prevents the prosecution, by their own undertaking, from adducing such evidence for the criminal court’s consideration in the first place at the threat of contempt. It is therefore possible to protect both the defendant and claimant’s interest without falling foul of Lord Wilberforce’s concern.

As to the majority’s concern in Reid, it is questionable whether it is really “inimical” to justice to order disclosure while limiting its use. It is not uncommon that the use of disclosed information in civil proceedings is limited. For instance, there’s the trite general rule that documents obtained during discovery must be used for the purpose of the litigation at hand only. It is therefore not necessarily against the interest of justice to limit the use of documents to the conduct of the civil litigation.

Statutory reform

However, is statute in the form of prevention of use the way to move forward? Of course it depends on the wording of a provision. But any statutory reform must address problems including but not limited to ones discussed in this short piece. Lord Neuberger MR in Gray28 and Professor Andrews29 supported statutory reform along the lines of section 72(3) of the Senior Courts Act 1981. The section applies in a passing off or infringement of IP rights action. The PSI was abrogated and replaced by a prohibition against its use in subsequent prosecution of related criminal offence30. Section 72(3) reads:

“…no statement or admission made by a person—

(a)in answering a question put to him in any proceedings to which subsection (1) applies; or

(b)in complying with any order made in any such proceedings, shall, in proceedings for any related offence or for the recovery of any related penalty, be admissible in evidence against that person…”

It is submitted that there may be two obvious problems for a section 72(3) approach. Firstly, there must be a clear supervision of the scope of disclosure. It should not be used by defendants to disclose evidence not necessarily required under the order. This may potentially become a haven for preventing the prosecution from relying on evidence that should’ve been admissible at trial. One way to do this may add a discretion to the trial judge to allow an evidence if it is clear that such evidence is not so covered under the order. The danger for the defendant is to only disclose what is necessarily required, this can be achieved if the court makes a clear order as to specific information or documents to be disclosed.

Secondly, there does not seem to be a residual discretion to disallow disclosure. Such discretion may be called for, for instance, when the CPS prosecution will likely collapse because of the precluded evidence. It is not unthinkable that the CPS may obtain the evidence from a different source before a disclosure order is made. The order would preclude the use of such independently obtained evidence from being adduced. The interest of the CPS would certainly be undermined by this broad-brush approach. Such a problem cannot be addressed by asking what one values more: catching the bad guys or protecting the innocent. It is a policy consideration that must be undertaken by the legislature.

It is submitted that statutory reform is, however, preferable to existing common law approach. Firstly, it addresses the reluctance of some judges in limiting the common law PSI right. This benefit is perhaps less important in England & Wales since the highest court has not shared such concern. Secondly, and perhaps more importantly, statutory reform would solve the practicality objection. The CPS will no longer be the de facto arbiter of whether a claimant gets a relief where the defendant asserts PSI. It also means that the legal certainty objection is also met. Of course, this is assuming Parliament reforms the law along the lines of section 72(3). Secondly, it addresses the problem of the E&W approach where the CPS retains a discretion which ultimately affects whether the claimant gets the disclosure order. As noted above, the court cannot bind the CPS unless the CPS gives a voluntary undertaking to court. Such that the effect of the E&W approach is not so different to the Australian one since the CPS has little incentive to give such undertaking. A statutory provision would resolve this current discrepancy between principle and practice.

Taking stock from the foregoing discussion, it is hoped the readers are at least persuaded the current common law approaches (both E&W and Australia) are problematic and needs urgent reform. The PSI deserves protection by our legal system, yet it should be relied on to categorically undermine the effectiveness of civil process. Since courts cannot bind the prosecution without its consent, statutory reform is needed to ensure the aim of PSI is served: to prevent use of disclosed information in subsequent criminal prosecution of the defendant for a related offence. It is prime time for this reform.

[1] Prof. John Sorabji, “Calling time on the privilege against self-incrimination in civil proceedings?” (2012) 31 CJQ 261.

[2] Ibid.

[3] 378 US 52 (1964).

[4] Ibid.

[5] The name came from Mareva Cia Naviera SA v international Bulk Carriers [1980] 1 ALL ER 213.

[6] See for example Vasil v National Australia Bank Ltd 46 NSWLR 207.

[7] Gray v News Group Newspaper Ltd [2012] EWCA Civ 48, at [17].

[8] AT&T Istel Ltd v Tully [1993] AC 45 at 53B.

[9] Gray (n 7) at [18].

[10] American Cyanamid Co v Ethicon Ltd [1975] AC 396.

[11] Bishopsgate Investment Management Ltd & Ors v Maxwell & Ors [1993] Ch. 1 at 38E per Dillon LJ.

[12] Rank Film Distributors Ltd & Ors v Video Information Centre (A Firm) & Ors [1982] AC 380, at 443 per Lord Wilberforce.

[13] Named after Anton Piller KG v Manufacturing Processes Ltd [1976] Ch 55.

[14] Rank Film Distributors Ltd (n 12).

[15] United Norwest Co-Operatives Ltd v Johnstone & Ors (unrep, The Times, 24 February 1994).

[16] AT&T Istel Ltd (n 8).

[17] Ibid 63F-64A (Lord Ackner).

[18] Reid v Howard (1993) 31 NSWLR 298 at 308-309.

[19] AT&T Istel Ltd (n 8).

[20] Reid v Howard (1995) 184 CLR 1 at 11, per Toohey, Gaudron, McHugh and Gummow JJ.

[21] Vasil (n 6) [24].

[22] Reid (n 20).

[23] Neil Andrews, “Privilege against self-incrimination and civil proceedings” (1993) 52 CLJ 42 at 44.

[24] AT&T (n 8).

[25] Reid (n 20) [14].

[26] Rank Film Distributors Ltd (n 12) 442F-G.

[27] AT&T (n 8).

[28] Gray (n 7) [18] (Lord Neuberger MR).

[29] Neil Andrews (n 23).

[30] [30] With the exception of perjury and contempt.

Article by Jason Lin
LLB (London School of Economics and Political Science) '21 and Criminal Law Notes Editor of the LSE Law Review Summer Board 2019